Integrating truncated passives in creole

PSYCHOLOGY 5-014
HUMAN LEARNING AND MEMORY


AN INVESTIGATION OF CONTEXTUAL EFFECTS ON THE
RECOGNITION MEMORY FOR TRUNCATED PASSIVES




ROBERT K. PHILIPS
9 JULY 1979

INTRODUCTION

In the teaching and learning of a second language, the presence of common structural features can be utilised to develop facile instructional procedures. However, where the first and second languages are dialects derived from a common language, Jamaican Creole and Standard English for example, the related structural features can interfere with each other and make the learning of the target language more difficult. Specifically, it has been claimed that "[t]he Creole verb does not have a distinct passive form (Bailey, 1966, p. 146). Cn the surface, there should be no structural to the learning of the English passive by Creole speakers, but there is evidence that Creole speakers comprehend and produce a form of passive in which there is no copula e.g. "The house sell" which is equivalent in meaning to "the house is (was) sold". (Edwards, 1978) There being no prior research on the memory for sentences among Creole speakers, it will be necessary to examine the literature for the differences between active and passive English sentences, and on truncated passives in particular.

REVIEW

The sentence "the man hit the ball" is an active sentence with the order logical subject-verb-logical object. "The ball was hit by the man" is the corresponding passive. Note that logical subject and logical object correspond to grammatical subject and object respectively in the active, but not in the passive. In a propositional formalism similar to that used by Kintsch (1974), the active can be represented by the relation R and ordered pair (x,y), so "the man hit the ball" is R(x,y) and "the ball hit the man" is R(y,x). Hence R(x,y) is not equivalent to R(y,x).

Clifton and Odom (1966) [cited in Clark (1976)] showed that simple active sentences and their corresponding passives were judged to be more similar than any other pair of negative, negative passive, passive question, negative question and passive negative question. In terms of the formalism introduced above, the passive sentence "the ball was hit by the man" could be represented by R~1 (y,x). So R(x,y) = R~1 (y,x).
Some passives are non-reversible (Slobin, 1966). For example, in the sentence "the burglar was arrested by the policeman", there is no need to specify the order information in the pair of R (b,p) since it is not expected that burglars arrest policemen. More clearly, for the sentence "the flower was picked by the girl", R (f,g) can have one meaning only, since one of the nominals is inanimate.

The "agentless"' or truncated passive (Slobin, 1968) is one in which the logical subject is deleted. For instance, "the floor was washed" is agentless. Such a sentence contrasts with "the floor was washed by the janitor." (Franks and Bransford, 1974). The Creole sentence, "the house sell" whose equivalent was given earlier, is a form of truncated passive. According to Anisfeld and Klenbort (1973), the agentless passive is the more frequently used form, with the passive in general being "used when the logical subject is unknown, cannot be easily stated, or is self-evident," (p. 121).

Clark (1976) claims that an active sentence and its corresponding passive, while being similar in expressing the same underlying proposition, are different in what they assert. "Fats killed the roach" presupposes that "Fats did something", whereas "The roach was killed by Fats"t presupposes that "something happened to the roach". Clark used Wright's (1969) data on question answering of active and passive sentences to justify his analysis since questions and sentences with the same presuppositions gave fewer errors than those with different presuppositions. Anisfeld and Klenbort (1973) offer a similar analysis, with the further claim that if presupposed information alone is being given, the truncated passive is used.
Slobin (1968) demonstrated that Ss tended to recall verbatim, truncated passives to a greater extent than full passives. Franks and Bransford (1974) read three paragraphs to subjects, in which there were truncated passives, full passives and actives, and one sentence with a generalized actor. For one of the truncated passives, "the actor was supplied by the immediately succeeding sentence".

Franks and Bransford hypothesised that Ss would "recognize" sentences that could be inferred from the truncated passive and the succeeding sentence, but had not been present in the paragraphs. Their hypothesis was confirmed in that 74% of the truncated passives with no actor were recognized, but only 39% of the truncated passives with an actor (SP-semantics) were recognized as truncated passives. A result that was evidently less interesting to Franks and Bransford was that 34% and 27% of the SP-semantics were recognized as full passives and full actives respectively.

In a recall task, Barclay and Reid (1974) observed the integration of truncated passives/actor to full actives and passives with kindergarten, first, third and fifth graders. Perfetti and Goldman (1975) gave Ss a prompted recall task in which the to be recalled sentence was either active or passive, and the sentence was first encountered in a paragraph whose theme was varied. "Theme" was defined as "which of two nouns from the text sentence.. had its referent more frequently involved in the preceding part of the narrative passages." The topic of the sentence was simply the noun that was first in it. The probe was agent-topic, recipient-topic, agent-comment or recipient-comment. It was found that the logical subject (agent) was a better prompt than the object (recipient); a thematized noun was a better prompt than a non-thematized noun. "Recoding" from the passive to active was possible where agent-comment and recipient-topic probes were used, and took place 85% of cases in the immediate recall condition, with thematized prompts producing more recoding than non-thematized prompts, 90% vs 80%. Perfetti and Goldman contend that a theme provides "an organization for memory", so that the thematized noun is used as the subject of an active sentence, on recall. It is likely that on a recognition task, the presence of a theme prior to a truncated passive sentence will bias the recognition of an active rather than a passive sentence.

Finally, Edwards (1978) constructed a sub-test of five items on a Creole Interference Test, in which an active sentence in British English was given with three alternatives, e.g.
Impalas eat frequently at this time of the year.
a) Yes, they taste nice.
b) Yes, they eat a lot of food.
c) Yes, they give a lot to eat.
In this item the intention was to discriminate between Ss who interpret the sentence as "Impalas are eaten frequently at this time of the year," and those with a Standard English interpretation. An unfamiliar noun was chosen to limit the judgment to that used for reversible passives. Scores on this subtest were not reported, but significant differences were obtained between British and West Indian children (ages 12 to 16). It is evident then that a passive interpretation of the verb can be induced.

HYPOTHESIS

It is contended here that it would be interesting to see whether the unmarked passive discussed above, can be integrated semantically in a manner similar to results obtained by Franks and Bransford (1974). Given the tendency to recall a passive as an active when the theme of the preceding discourse is the logical subject of the TBR sentence, (Perfetti & Goldman, 1975), it should be possible to bias the recognition of actives where truncated passives are presented on acquisition. Enhanced recognition of actives would be in accordance with the prior mention of the presupposition appropriate to an active sentence (Anisfeld & Klenbort, 1973). So "Mrs. Brown was a very good cook. After the harvest a hugh feast was served", could be integrated to the sentence. "Af ter the harvest Mrs. Brown served a huge feast."

Demonstration of the effect of the integration to actives in English, would provide a basis for showing the same effect in Creole sentences: for example "John Brown (is) not hungry. The dinner eat already" could be integrated to "John Brown eat dinner already". To reiterate the reason for the need to show integration to actives, there is no means for stating "Dinner has been eaten by John Brown already" using the passive in Creole. The attempt to demonstrate integration with Creole sentences will be contingent on the findings of the experiment, and will be pursued in a future study.

In summary, it is expected that where consistent prior reference is made to the actor, the truncated passive will be recognized as an active sentence, significantly more than when no such reference is made.

METHOD

Three paragraphs were read to two groups of subjects, who were asked to listen to them carefully in order to answer questions about them. For one group the paragraphs made either considerable prior reference or no prior reference to the agent in the target sentences. For each target sentence there was a block of five sentences, one of which was the target. Three minutes after hearing the last paragraph Ss were required to mark the sentences that they felt had been presented.
An example of a paragraph and a block of sentences is shown below. The company storeroom was hot and stuffy. The airconditioner was turned off. The janitor’s corner was cluttered with boxes. He was angry because he was perspiring. He opened his shirt and shook it rapidly. In the main office the accountant was asked if he felt hot. The janitor/clerk said that he was sorry to hear. All the typists stopped working. They too were hot. Cold drinks were brought in by someone. The workers were sent home by the union steward.

In the main office, the accountant was asked by the janitor/clerk if he felt hot.
In the main office, the accountant was asked if he felt hot.
In the main office, the janitor/clerk asked the accountant if he felt hot.
In the main office, someone asked the accountant if he felt hot.
In the main office, the accountant was asked by someone if he felt hot.

RESULTS AND CONCLUSIONS (not verified by testing)

RECOGNITION SENTENCE BLOCK
Acquisition Full Full Short Generalized
Sentence Passive Active Passive
Full Passive .45 .45 .10 .00
Short Passive
Alone .10 .10 .80 .00
Short Passive
Semantic .36 .28 .36 .00
Short Passive
Semantic Theme .20 .44 .36 .00





The effect of providing a theme by having prior sentences about the actor, should enhance the "recognition" to full actives; the proportion full actives recognized in the "prior reference" conditions should be significantly greater than in "no reference" condition .44 vs .28. On strength of this result it should be possible to construct paragraphs which Creole speakers can integrate the truncated passive to a full active. If this further result is obtained, then it can be held tentatively, that all is required for the Creole speaker in learning the English passive is the copula and the appropriate participle; surface structures rather than a "deep" cognitive rule. There would be some evidence that a structure that is functionally equivalent to the English passive is present.






















REFERENCES

Anisfeld, M., & Klenbort, I. On the functions of structural paraphrase: The view from the passive voice. Psychological Bulletin, 1973, 79,(2), ll7-126
Bailey, B. L. Jamaican Creole Syntax A Transformational Approach. Cambridge, Cambridge University Press, 1966.
Barclay, J. R. & Reid, M. Semantic Integration in Children’s Recall of Discourse. Developmental Psychology, 1974, 10,(2), 277-281.
Clark, H. H. Semantics and Comprehension. The Hague. Mouton, 1976.
Clifton, C., Jr. & Odom, P Similarity relations among certain English
sentence constructions. Psychological Monographs 80 (5, Whole No. 613) 1966.
Edwards, V. K. Dialect Interference in West Indian Children. Language & Speech, 1978, 76-86.
Franks, J. J., & Bransford, J. D. Memory for syntactic form as a function of semantic context. Journal of Experimental Psychology, 1974, 103,(5), 1037-1039.
Kintsch, W. The Representation of Meaning in Memory. Hillsdale, N.J. Erlbaum, 1974.
Perfetti, C. A., & Goldman, S. R. Discourse Functions of Thematization and Topicalization. Journal of Psycholinguistic Research, 1975, 4,(3), 257-271.
Slobin, D. I. Grammatical transformations and sentence comprehension in childhood and adulthood. Journal of Verbal Learning and Verbal Behavior, 1966, 5, 219-217.
Slobin, D. I. Recall of full and truncated passive sentences in connected discourse. Journal of Verbal Learning and Verbal Behavior, 1968, 7, 876-881.
Wright, P. Transformations and the understanding of sentences. Language & Speech,1969,12, 156-166